Update dpnp.linalg.matrix_rank() implementation

dpnp/backend/kernels/dpnp_krnl_linalg.cpp

@@ -579,15 +579,6 @@ template <typename _DataType>
 void (*dpnp_matrix_rank_default_c)(void *, void *, shape_elem_type *, size_t) =
     dpnp_matrix_rank_c<_DataType>;
 
-template <typename _DataType>
-DPCTLSyclEventRef (*dpnp_matrix_rank_ext_c)(DPCTLSyclQueueRef,
-                                            void *,
-                                            void *,
-                                            shape_elem_type *,
-                                            size_t,
-                                            const DPCTLEventVectorRef) =
-    dpnp_matrix_rank_c<_DataType>;
-
 template <typename _InputDT, typename _ComputeDT>
 DPCTLSyclEventRef dpnp_qr_c(DPCTLSyclQueueRef q_ref,
                             void *array1_in,

dpnp/linalg/dpnp_iface_linalg.py

@@ -424,6 +424,10 @@ def matrix_rank(A, tol=None, hermitian=False):
     rank : (...) dpnp.ndarray
         Rank of A.
 
+    See Also
+    --------
+    :obj:`dpnp.linalg.svd` : Singular Value Decomposition.
+
     Examples
     --------
     >>> import dpnp as np

dpnp/linalg/dpnp_utils_linalg.py

@@ -1014,20 +1014,12 @@ def dpnp_matrix_rank(A, tol=None, hermitian=False):
     S = dpnp_svd(A, compute_uv=False, hermitian=hermitian)
 
     if tol is None:
-        tol = (
-            S.max(axis=-1, keepdims=True)
-            * max(A.shape[-2:])
-            * dpnp.finfo(S.dtype).eps
-        )
+        rtol = max(A.shape[-2:]) * dpnp.finfo(S.dtype).eps
+        tol = S.max(axis=-1, keepdims=True) * rtol
     else:
         if dpnp.is_supported_array_type(tol):
-            # Check that `a` and `tol` are allocated on the same device
-            # and have the same queue. Otherwise, `ValueError`` will be raised.
-            get_usm_allocations([A, tol])
-        else:
-            # Allocate dpnp.ndarray if tol is a scalar
-            tol = dpnp.array(tol, usm_type=A.usm_type, sycl_queue=A.sycl_queue)
-        tol = tol[..., None]
+            # Add a new axis to match Numpy's output
+            tol = tol[..., None]
     return dpnp.count_nonzero(S > tol, axis=-1)
 
 

tests/test_linalg.py

@@ -566,7 +566,7 @@ def test_inv_errors(self):
 
 
 class TestMatrixRank:
-    @pytest.mark.parametrize("dtype", get_all_dtypes(no_bool=True))
+    @pytest.mark.parametrize("dtype", get_all_dtypes())
     @pytest.mark.parametrize(
         "data",
         [
@@ -586,7 +586,7 @@ def test_matrix_rank(self, data, dtype):
         dp_rank = inp.linalg.matrix_rank(a_dp)
         assert np_rank == dp_rank
 
-    @pytest.mark.parametrize("dtype", get_all_dtypes(no_bool=True))
+    @pytest.mark.parametrize("dtype", get_all_dtypes())
     @pytest.mark.parametrize(
         "data",
         [
@@ -666,7 +666,12 @@ def test_matrix_rank_errors(self):
         tol_queue = dpctl.SyclQueue()
         a_dp_q = inp.array(a_dp, sycl_queue=a_queue)
         tol_dp_q = inp.array([0.5], dtype="float32", sycl_queue=tol_queue)
-        assert_raises(ValueError, inp.linalg.matrix_rank, a_dp_q, tol_dp_q)
+        assert_raises(
+            dpctl.utils._compute_follows_data.ExecutionPlacementError,
+            inp.linalg.matrix_rank,
+            a_dp_q,
+            tol_dp_q,
+        )
 
 
 @pytest.mark.usefixtures("allow_fall_back_on_numpy")